2016-01-11 - production
18:30	<ori>	Restarting HHVM on all job runners, to vacate memory now that the cause of the leak appears to have subsided.(T122069)	[production]